1. The Burden of Work-related Cancer in Great Britain L Rushton 1, T. Brown 2, R Bevan 3, J Cherrie 4, L Fortunato 1, S Bagga 3, P Holmes 3, S Hutchings 1, R Slack 3, M Van Tongeren 4, C Young 2 1 Dept. of Epidemiology and Public Health, Imperial College London; 2 Health and Safety Laboratory, Buxton, Derbyshire 3 Institute of Environment and Health, Cranfield University 4 Institute of Occupational Medicine This study was funded by the Health and Safety Executive

2. Aims of the overall study Current Burden of Occupational Cancer: –Estimate size of current burden based on past exposures at work –to identify important cancer sites –to identify industries and occupations for targeting for reduction measures –Estimation carried out for all substances and circumstances (e.g. work as a painter or welder) in the workplace defined by International Agency for Research on Cancer as definite (group 1) and probable (group 2A) human carcinogens Prediction of Future Burden of Occupational Cancer –Estimate size of future burden based on current and past exposures –Identify cancer sites, carcinogens and industry sectors where the burden is greatest –Demonstrate effects of measures to reduce exposure

3. Methods Measure of burden: Attributable Fraction (AF) - proportion of cases attributable to exposure; needs –risk of disease associated with the exposure of concern: obtained from relevant published literature –proportion exposed in the population To take into account latency (length of time before disease risk increases) we defined the risk exposure period (REP) for: –Solid tumours: 10-50 years; 1956-95 –Leukaemia: up to 20 years; 1986-2005 Proportion exposed over the REP is: number ever exposed/number ever worked Estimated using national data sources (CAREX, LFS, CoE) Adjusted for turnover, change in numbers employed over REP e.g. Manufacturing decreasing, service sector increasing

4. Cancer site:Attributable Fraction(%)Attrib Deaths (2005)Attrib. Registrations (2004) MaleFemaleTotalMaleFemaleTotalMaleFemaleTotal Bladder7.11.95.32153024549654550 Bone0.0 000000 Brain0.50.10.31011112214 Breast 4.6 555 1,969 Cervix 0.7 77 18 Kidney0.04 111213 Larynx2.91.62.61732050656 Leukaemia0.90.50.71852330938 Liver0.20.10.2425415 Lung21.15.314.54,0207254,7454,6278155,442 Lympho-haematopoietic0.0040.0020.003000000 Melanoma (eye)2.90.41.6101616 Mesothelioma97.082.594.91,6992381,9371,6992381,937 Multiple Myeloma0.40.10.35168210 Nasopharynx10.82.48.071814115 NHL2.11.11.743145710239140 NMSC6.91.14.5202232,5133492,862 Oesophagus3.31.12.51562818415929188 Ovary 0.5 23 33 Pancreas0.020.01 101101 Sinonasal43.319.832.72710389531126 Soft Tissue Sarcoma3.41.12.41131322427 Stomach3.00.31.910161081499157 Thyroid0.120.020.05000101 Total8.22.35.36,3551,6558,0109,9883,61113,598 Total GB cancers 15+yrs 77,91272,212150,124175,399168,184343,583

5. Cancer SiteAsbestosShift work Min. oils Solar rad n SilicaDEEPAHs (Tars) PaintersDioxinsETSRadonWeldersAll Bladder 296 106 71 550 Brain 14 Breast 1,957 1,969 Cervix 18 Kidney 3 Larynx8 56 Leukaemia 38 Liver 5 Lung2,223 470 907695 282215284209175 5,442 LH cancers 1 Melanoma eye 6 Mesothelioma1,937 Multiple Myeloma 10 Nasopharynx 15 NHL 74 140 NMSC 9021,541 475 2,862 Oesophagus 188 Ovary 33 Pancreas 1 Sinonasal 55 126 STS 27 Stomach47 83 157 Thyroid 1 Total Attrib. Registrations 4,2161,9571,7221,54190780147543731628420917513,598

6. Arsenic Asbestos Beryllium Cadmium Chromium IV Cobalt Diesel engine exhaust ETS Inorganic lead Ionising radiation Mineral oils Nickel PAHs Painters Radon Silica Steel foundry workers Strong inorganic-acid mists TCDD (Dioxins) Tin miners Welders Lung cancer AF by carcinogen/occupation

7. IndustryAsbestosShift work Min. oils Solar rad n SilicaDEEPAHs (tars) PaintersDioxinsETSAll Total Agriculture, farming 135 55 263 Iron and steel basic industries 00 04 75 135 Manufacture industrial chemicals64 11 11 121 Manufacture of instruments, photographic and optical goods 203 206 Manufacture machinery not electrical 0 82 111 Manufacture of other chemical products 69 101 123 Manufacture transport equipment115 05112 188 Metal workers 1,252 1,250 Mining197 312943 302 Non-ferrous metal industries 942 50 159 Painters (not construction) 102 Printing, publishing et c 2673 0 286 Welders 182 Total manufacturing etc 535 1,722163200804102254 3,944 Construction 2,773 841707290 364,816 Painters/decorators (construction) 334 335 Roofers/road workers (construction) 471 541 Total Construction 2,773 841707290471334 36 5439 Land transport 133 6 350 3505 Personal/household services 361 714 29 22804 Public admin./defence 240 20273 Shift work 1,957 Wholesale, retail, restaurants 66 6 6 7118269 Total service industry 5731,9577402 431 72484,177 Total Attrib. Registrations4,2161,9571,7221,54190780147543731628413,598

8. Cancer Registrations Attributable to Work in the Construction Industry - Men 05001000150020002500300035004000 Arsenic Asbestos Chromium Cobalt Diesel ETS Formaldehyde Lead Painters PAH PAH - coal tars and pitches Radon Silica Solar Radiation Tetrachloroethylene Wood dust Carcinogen Number of Registrations Construction, inc painters and decorators; road surfacers, roadmen, roofers & glazers, paviours Other Sectors

9. Predicting Future Burden Attributable Fractions and attributable numbers of deaths and cancer registrations estimated for a series of forecast years, e.g. 2010, 2020 … 2060 Define the risk exposure period (REP) for each year e.g. for 2030, 1981 – 2020 Changing balance between past and future exposure Forecasted AFs take into account employment turnover and changes in different industry sector employment trends Method developed to shift the proportion of workers exposed in different exposure level categories (H/M/L/B) across time as exposures gradually decrease Predicted numbers based on demographic change only i.e. Assuming all non-occupational risk factors e.g. Smoking stay same as 2004/5 Method provides a tool for comparing ‘doing nothing’ (baseline scenario) with various interventions Methods applied to top 14 carcinogens/occupations identified as accounting for 85.7% of total current (2004) cancer registrations

10. Change in future exposure: Intervention Scenarios Can test: Introduction of a range of possible exposure standards or reduction of a current exposure limit Improved compliance to an existing exposure standard Comparison of lowering an exposure standard versus improved compliance Planned intervention such as engineering controls or introduction of personal protective equipment Industry closure Also can vary: Timing of introduction (2010, 2020 etc) Compliance levels e.g. according to workplace size (self- employed, 1-49, 50-249, 250+ employees) Intervention scenario results compared to the ‘baseline – no change’ or ‘baseline - trend’ scenario to assess relative impact on reducing attributable numbers

11. No appropriate exposure measurements ETS (lung cancer) – test compliance to smoking bans Radon (lung cancer) – reduce exposed nos. by 10% per decade Solar radiation (NMSC) – move workers into lower exposure (time spent outdoors) categories Occupational circumstances, no specified ‘carcinogen’ Shift work (breast cancer) - move workers into lower duration of exposure categories Painters, welders – reduce excess risk Some exposure data – standards can be tested RCS (lung cancer) – test existing and stricter standards, estimate current compliance and test effect of variable compliance Arsenic, strong acids, tetrachloroethylene, TCDD – test possible standards based on H/L exposure boundary estimates, or L/B below which excess risk is zero DEE – test suggested OEL Intervention Scenarios for Priority Carcinogens

12. Carcinogens where occupational standards/limits exist or could be introduced Example: silica –Reduce exposure limit from 0.1 to 0.05 mg/m 3 in all workplaces, in 2010, with the same proportion exposed above the new limit as above the old –Reduce exposure limit again to 0.025 mg/m 3 –Improve compliance from 33% to 90% in all workplaces –Try doing both for all workplaces –Successively enforce the new limit and improve compliance in workplaces of different sizes

13. Testing reduction of exposure standard and changes in compliance Forecast lung cancers for 2060 for Respirable Crystalline Silica 2010 Attributable Fraction Attributable registrations Avoided registrations 3.3803 2060 Base-line: exposure limit 0.1mg/m 3, compliance 33% 1.08794 Exposure limit 0.05mg/m 3, compliance 33% 0.80592202 Exposure limit 0.025mg/m 3, compliance 33% 0.56409385 Exposure limit 0.1mg/m 3, compliance 90% 0.14102693 Exposure limit 0.05mg/m 3, compliance 90% 0.0749745 Exposure limit 0.025mg/m 3, compliance 90% 0.0321773

14. Lung cancer from exposure to RCS Effect of reducing the exposure standard for RCS versus compliance Attributable registrationsAFs

15. Testing improvement in compliance by workplace size Forecast lung cancers for 2060 for Respirable Crystalline Silica 2010 Attributable Fraction % Attributable registrations Avoided registrations 3.3803 2060 Base-line: exposure limit 0.1mg/m 3, compliance 33% 1.08794 Exposure limit 0.05mg/m 3, compliance 33% 0.80592202 Exposure limit 0.05mg/m 3, % compliance changes by employed workplace size and self employed 33% < 250, self employed; 90% 250+ 0.68499295 33% < 50, self employed; 90% 50+ 0.61451344 33% self employed; 90% all sizes employed 0.35261533 90% all workplaces 0.0749745

16. Occupational Circumstances no ‘exposure data’ Example: Shift Work (Night work) Breast cancer: important contribution to the total current occupational cancer burden Exposure defined by nature of occupation – unknown agent, no exposure data Evidence of dose response with duration of night work Duration Relative Risk Proportion ‘exposed’ <5 years: 0.9530% 5-14 years: 1.2940% 15+ years:2.2130% Intervention scenarios expressed as limiting proportions in night work for durations of 15+ and 5+ years

17. Shift (Night) Work: Attributable Cancers (1)Current employment levels maintained, 30% <5, 40% 5- 14, 30% 15+ years night shift work (2)Linear employment trends to 2021-30 (3)50%<5, 30% 5-14, 20% 15+ years night shift work (4)70%<5, 20% 5-14, 10% 15+ (5)90%<5, 10% 5-14, 0% 15+ (6)100% <5 years

18. Summary of Future Burden Results 14 agents account for 85.7% current occupation attributable cancer (2004), 12,000 cancers in 2010 Will rise to nearly 13,000 by 2060 given current trends in employment and exposure levels (>12,300 if current levels maintained). Aging population is a factor. No impact seen until 2030 because of general increase in cancers due to aging population With modest intervention over 2,000 cancers can be avoided by 2060 (including 376 lung, 928 breast cancers, 432 NMSC) With stronger interventions nearly 8,500 can be avoided by 2060 (including 1,732 lung, 3,062 breast and 3,287 NMSC) Methods enables effective interventions to be identified Need to monitor exposure levels in future to assess whether interventions have been successful

19. Uncertainties and the impact on the burden estimation Source of UncertaintyPotential impact on burden estimate Exclusion of IARC group 2B and unknown carcinogens e.g. for electrical workers and leukaemia ↓ Inappropriate choice of source study for risk estimate ↑↓ Imprecision in source risk estimate ↑↓ Source risk estimate from study of highly exposed workers applied to lower exposed target population ↑ Risk estimate biased down by healthy worker effect, exposure misclassification in both study and reference population ↓ Inaccurate latency/risk exposure period, e.g. most recent 20 years used for leukaemia, up to 50 years solid tumours ↓ Effect of unmeasured confounders ↑↓ Unknown proportion exposed at different levels ↑↓

20. Summary Robust methodological approach developed for estimation of burden for all IARC 1, 2A occupational carcinogens Outputs: Current burden results –Preliminary work on 6 cancer sites: Occupational and Environmental Medicine 2008, 65, 789-800; –Results from all sites: Br J Cancer 2010, 102: 1428-1437 + Technical report on HSE website –Supplement (13 papers) of current burden detailed results Br J Cancer 2012;107(S1):S1-S108 23 technical reports available at http://www.hse.gov.uk/cancer/ –Papers in preparation reporting –DALYs/inequality of burden –Evaluation of impact of source of bias and uncertainty Outputs: Future burden –Methodology paper: Am J Epidem 2011, 173, 1069-1077+ technical report on HSE website –Future burden results: in press Cancer Prevention Research